'Department of Biochemistry, McMaster University, Hamilton,. Ontario, Canada ...... amino acid match or mis-match. The scores are based on the ... 36 itm mesh, and analyzed with a Coulter ModelEpics-C Flow Cytometer. To quantify the GI ...
The EMBO Journal vol.7 no.13 pp.4335-4346, 1988
The WHI 1 + gene of Saccharomyces cerevisiae tethers cell division to cell size and is a cyclin homolog
Robert Nash, George Tokiwa, Sukhjit Anand1, Kristen Erickson and A.Bruce Futcher PO Box 100, Cold Spring Harbor Laboratory, NY 11724, USA and 'Department of Biochemistry, McMaster University, Hamilton, Ontario, Canada Communicated by J.D.Watson
WHII-I is a dominant mutation that reduces cell volume by allowing cells to commit to division at abnormally small sizes, shortening the GI phase of the cell cycle. The gene was cloned, and dosage studies indicated that the normal gene activated commitment to division in a dose-dependent manner, and that the mutant gene had a hyperactive but qualitatively similar function. Mild over-expression of the mutant gene eliminated GI phase, apparently entirely relaxing the normal GI size control, but revealing hitherto cryptic controls. Sequence analysis showed that the hyperactivity of the mutant was caused by the loss of the C-terminal third of the wild-type protein. This portion of the protein contained PEST regions, which may be signals for protein degradation. The WHll protein had sequence similarity to clam cyclin A, to sea urchin cyclin and to Schizosaccharomyces pombe cdc13, a cyclin homolog. Since cyclins are inducers of mitosis, WHIl may be a direct regulator of commitment to division. A probable accessory function of the WHIl activator is to assist recovery from a factor arrest; WHII-I mutant cells could not be permanently arrested by pheromone, consistent with a hyperactivation of division. Key words: cell cycle/cyclin/cell size/PEST hypothesis/ WHII
Introduction Coordination between growth and division is essential for all cells. This is obvious for microbes, where growth is often the sole requirement for division, but multi-cellular eukaryotes also have a basic requirement for cellular growth underlying a more complex network of signals regulating cell division. Most eukaryotic cells commit themselves to division at a particular control point, variously called the commitment point, the restriction point or Start (Pringle and Hartwell, 1981; Pardee et al., 1978). In many organisms, including Saccharomyces cerevisiae (Johnston et al., 1977), Schizosaccharomyces pombe (Fantes, 1977), Chlamydomonas reinhardii (Donnan and John, 1983), Amoeba proteus (Prescott, 1956) and some mammalian cells (Killander and Zetterberg, 1965; Yen et al., 1985; Shields et al., 1978; MacQueen and Johnson, 1983), it has been shown that growth to a critical cell volume is a necessary prerequisite for commitment. Once committed, cells inevitably complete division, and arrive back at the ©IRL Press Limited, Oxford, England
commitment point ready to consider another round of division. Our goal has been to discover the molecular nature of this commitment event. To this end, we have studied mutations that change the time of commitment. In the yeasts S. cerevisiae and S.pombe, growth to critical size is the primary control on division. In S.pombe, for instance, division of fast-growing cells is regulated primarily at the G2/M boundary by a size requirement for commitment to mitosis (slow-growing cells have an additional GI size control) (Nurse, 1975; Fantes, 1977; Nurse and Thuriaux, 1977). The wee] - mutation reduces the critical size required for mitosis, with the net result that wee] cells are only half the volume of wild-type cells (Nurse, 1975; Nurse and Thuriaux, 1980). The wee] + gene has been cloned, and sequence analysis suggests that it encodes a protein kinase (Russell and Nurse, 1987). When multiple copies of wee] + are chromosomally integrated, cell volume increases directly in proportion to copy number (Russell and Nurse, 1987). This suggests wee] + is a dosage-dependent inhibitor of mitosis. It may be that the cell is titrating a constant amount of wee I' protein against an increasing volume, and only when cell growth has diluted it to a sufficiently low level can mitosis occur. If this view is correct, then wee 1 may be the metric by which the proper time for mitosis is determined. In contrast, S. cerevisiae has a cell cycle controlled mainly by a GI commitment point called Start (reviewed by Pringle and Hartwell, 1981). Attainment of a critical size is a prerequisite for Start (Johnston et al., 1977; Hartwell and Unger, 1977; Jagadish and Carter, 1977; Lord and Wheals, 1980). Although many cell division cycle (cdc) genes have been identified that are required for Start (Hartwell et al., 1973; Reed, 1980), there is no evidence that these genes regulate the time of Start; that is, they may be part of the machinery for implementing Start, as opposed to regulating it. However, a non-lethal mutation called WHIll- is known to affect the time of Start; WHiII-I cells can initiate new cell cycles at half the volume of wild-type (i.e. WHII +) cells, and so are about half the volume of wild-type cells during exponential growth (Carter and Sudbery, 1980; Sudbery et al., 1980). Furthermore, like wee], WHIl-l is a co-dominant mutation, in that heterozygous diploids are intermediate in volume between homozygous mutants and homozygous wild-types (Sudbery et al., 1980). This suggests that the dosage of the WHI1-l protein is important, making it a good candidate for a molecule that might be used by the cell to titrate volume or some other signal connected with growth. Since WHII-I acts in GI (rather than at mitosis, as weel + does), we anticipated that it might play a relatively direct role in regulating cell division. (A note on nomenclature: the mutation was originally named whil-] with the expectation that it was a null mutation, and that its co-dominance was due to a dosage effect. However, we show below that the mutation is truly dominant, and so we have taken the liberty of re-naming 4335
R.Nash et al.
mapped WHI]-] to the left arm of chromosome I. CDC24 and WHII-] were very tightly linked (0 recombinants in 22 tetrads). Kaback, Pringle and co-workers (e.g. Coleman et al., 1986) mapped the transcription units near CDC24, calling transcription units without a known function FUN genes (Function Unknown Now). Our genetic data suggested that WHII-] was in FUN9, FUNIO or CDC24 (Figure 2). While cloning these regions, we learned from Dr F.Cross (personal communication) that he had isolated a new mutation, DAF], and mapped it to FUNJO. DAF] and WHI]-] had very similar phenotypes, and so it seemed likely that they were
it WHI]-1, in accordance with standard S. cerevisiae nomenclature. We refer to the wild-type allele as WHI] . Deletion alleles are referred to in lower case, e.g. whil-310. The protein is referred to without italics, e.g. WHII.)
Results Basic phenotypes of WHIl-1 The WHII-1 strain S673a was crossed to strain LL20, and cell size in the spore clones was assayed with a Coulter Channelyzer. After three backcrosses, the small cell phenotype always segregated 2:2 (Figure 1). The growth rates of WHII-I and WHII + cells were indistinguishable (data not shown). When glucose was replaced by glycerol as the carbon source, both cell types grew more slowly, and both cell types had reduced volumes (Figure 1). At low growth rates wild-type cells had a biphasic size distribution, because daughter cells were smaller than mother cells (previously budded cells). WHII-1 cells had a monophasic distribution because they divided relatively symmetrically at both high and low growth rates; the significance of this is not known.
allelic. An ApaI- Hpa I fragment encompassing FUN] 0 and 40 amino acids of CYC3 was therefore cloned from a WHII-I strain (Materials and methods). When this DNA (plasmid pBF30-Figure 2) was used to replace the FUNJ0 locus of a wild-type strain, the strain became Whi-. The same restriction fragment was also cloned from a wild-type strain and, when it was used to replace the DNA of a WHII-I strain, the strain became Whi+ (i.e. wild-type). An experiment described below shows definitively that WHI]-1 is in FUNJO rather than in CYC3. Both strands of FUNJO were sequenced (Figure 3). There is an open reading frame of 1740 nucleotides potentially
Mapping, cloning and sequencing WHIl- 1 Using a combination of classical mapping techniques and transposon tagging (method to be published elsewhere), we 1000
0) (1) 0 Jo
E z
0
ro
E
(D _
rio
E
rO\
0
4'
0
12
24
36 48 60 Cell Volume (,4m3)
Fig. 1. Coulter Channelyzer plots of cell volume distributions. Panel A is tetrad BF328-4 segregating 2:2 for WHIJ-J; B is tetrad BF334-2. Panel C shows BF328-4a (WHII +) and BF3284c (WHII-J) growing in YEPG (glycerol medium). td is the culture doubling time. The modes of the peaks are indicated in panel C.
4336
Control of division by cell size
encoding a protein of 580 amino acids. The 5' end of the mRNA has not been mapped, but there is circumstantial evidence that the 5'-most AUG is the initiator codon. First, between the 5'-most AUG and the next in-frame AUG are 105 sense codons, a statistically improbable length for random sequence. Second, the 5'-most AUG is embedded in the sequence ACGAUGGC, which is very similar to the higher eukaryotic translation initiation consensus ACCAUGGN (Kozak, 1986), and is also similar to the consensus A(A or C)AAUGNC for yeast mRNAs (Hamilton et al., 1987). Both strands of FUNI O from a WHII-I mutant were also sequenced. The only difference found was a C to T transition at base 1210 of the open reading frame, which changes a CAG codon for glutamine to a UAG stop codon (Figure 3). This stop codon removes 177 amino acids from the carboxy terminus, but leaves the first 403 amino acids intact. Dr F. Cross (personal communication) has sequenced a different dominant WHII mutation (DAFI), and it is also a stop codon near base 1200. Computer searches revealed a similarity to three cyclinssea urchin cyclin (Pines and Hunt, 1987), clam cyclin A (Swenson et al., 1986) and S.pombe cdc13 (Booher and Beach, 1988; R.Booher, personal communication; B.Byers, WHI-I? Yes
Yes HRR
R
-
kb
-I
Bg Hp RH
X
S A
R
Bg
I
v\\
FUN 9
,>
I
CDC24
WHI (FUN 10) I...
CYC 3
PYK
.I Deleted in pWJ3IO
CYC3
P,Sp,H pUCI8
pUC18
Fig. 2. Restriction map of part of chromosome I and plasmids pWJ310 and pBF30. The orientation of FUN9 is not known. Two fragments that can transfer WHII-I are shown. The deletion/disruption allele whil-310 has the EcoRI-SalI chromosomal fragment replaced by the URA3 gene, thus deleting most of the WHI] gene; however, the last 500 bp of the gene are retained. Other deletion/disruption alleles have the same phenotype. Restriction sites: A: ApaI, Bg: BglII, B/Bg: BamHI/BglII hybrid, H: HindIII, Hp: HpaI, K: KpnI, P: PstI, R: EcoRI, S: Sall, Sc: SacI, Sm: SmnaI, Sp: SphI, SIX: SalIlXhoI hybrid, X: XhoI. Xb: XbaI, X(R): an EcoRI site converted to an XhoI site, X(S): a SailI site converted to an XhoI site. The XbaI site in CYC3 is not shown in the drawing of the chromosome. Some restriction site information was adapted from Coleman et al. (1986), from Miyamoto et al. (1987) and from Gallay and Rothstein (personal communication).
personal communication). The overall percentage identity beween WHIl and any one of the three cyclins is low; however, the percentage identity between any two of the cyclins is also rather low. The observation suggesting a family of proteins is that the three cyclins share certain small, highly conserved regions in the central parts of the proteins, and many of these regions are also found in WHI1. The region of greatest similarity is a segment of 100 amino acids stretching from residue 106 to 206 of WHI1 (KMR . . . WSI). In this segment, the three cyclins are 80-90% similar to each other (including conserved changes as well as identities), and WHII is just over 50% similar to the cyclin consensus (including conserved changes). When WHI1 was compared to each of the other three cyclins using only this core 100 amino acids, the IALIGN program of the Protein Identification Resource (Georgetown University, Washington, DC) gave scores that were 9-12 standard deviations from the mean, varying only slightly from cyclin to cyclin, and varying only slightly with different IALIGN parameters. Pairwise comparisons using the complete sequences were also done for us by G.Otto, using a more sophisticated statistical approach (Otto, 1986). According to this analysis, the similarities were significant at the level of 1.5 x 10-8 for WHII versus urchin cyclin, 4 x 10-9 for cdc13, and 6 x 10-"1 for clam cyclin A. With this statistical backing, we are confident that WHII is a cyclin homolog, even though it is clearly less closely related to the three other cyclins than they are to each other. An alignment of the four proteins is shown in Figure 4. Another feature of the four proteins is that they contain regions unusually rich in proline (P), glutamate (E), serine (S), threonine (T) and/or aspartate (D). It has been proposed that small regions rich in these amino acids and flanked by basic residues are specific signals for proteolysis (the PEST hypothesis-Rogers et al., 1986). The PEST region in clam cyclin A has been noted previously (Swenson et al., 1986). Clam and sea urchin cyclin are known to be catastrophically destroyed at mitosis, and so the idea that they contain a signal for degradation is particularly plausible. In Figure 4, the various PEST regions are underlined. Sea urchin cyclin has a relatively poor PEST region, unbounded by basic residues, but the other three proteins each have very good PEST regions. Each protein has one PEST region immediately N-terminal to the first highly conserved motif (SEY). In addition, cdc13 has multiple N-terminal PEST regions, and WHI1 has multiple C-terminal PEST regions. These latter PESTs are removed by the WHIJ-] mutation. A number of motifs that could allow processing of WHI1 were found; for instance, there were five occurrences of Lys Arg, two each of Arg Arg and Arg Lys and one of Lys Lys. These dibasics could allow proteolytic cleavage, and were also relatively abundant in the other three cyclins. Also, WHII contained eight possible glycosylation sites (Asn X Ser/Thr). Sea urchin cyclin and cdc13 also contain potential glycosylation sites. We do not yet know whether any of these potential processing sites is used. The codon bias of WHIJ + is 0.27, which suggests that the protein is not abundant (Bennetzen and Hall, 1982). WHI 1-1 is fully dominant, and is in the FUN 10 transcription unit A single copy of pBF30 (which carries WHII-]) was integrated at the XhoI site of the FUNJO locus of a wild-type 4337
R.Nash
et al.
CTGACAGAGA CACCCCTAGA GOCTACATTA CTGATTTCGG AAATTTCCCA AATTGGAAAT ATCACCTCGT
-931
311 CGACGTGCTG CGGTGCATOG CTGTTTTACC CGTTTAGGAA AAAACTCGGC GGGTTTTCTT GACGGOCAAA
-861
TGSCGCCATT CGTCTCGTTT GAACGCTTGG CCTGTAAAAC AGCTAACTCA TTCACTATCT CTATCCTGCG
-791
ACCCTGATGA TACCGAAMT ATGCTGATAC AGATCGGTCC CTCATATCGC GATAGGGCTT TCTGAGCTCT
-721
CCTCCCCCTT CTTCCCTCCC CAGCCAAAGA GGCGGTTTT CCTTCTOGGG AAGTGTCGCA ACCAAACGCT
-651
GTOCTCAAAC CCTAMATTGT GAACTTTCAA GAAAAMAAA AAGAAAAAGT GAAAAATTAT CAGGCAAGAA -581 AAAGAAATTA CCAAGCCTGC TCTCACTGTA ATGATCAAGT TACATAAATT TACTATCGGA TTAGTGTOTC
-511
GOT OCA ATT AAC CTC ATC AAA TTA TCT TTG AAC TAC TAT AAT TCA AAC CTT TOG GAA Gly Ala le Asn Leu IIl Lys Leu Sor Lou Asn Tyr Tyr Asn Sbr Ann Lou Trp Glu
954
AAT ATC AAT CTG GCT TTG GAG GAA AAC TOC CA GAC CTA OAT ATT AAA TT$ TCA GAA Ann Ie* Amn Leu Ala Lou Glu Glu Asn Cys Gln Amp Lou Asp Ill Lys Lou Sr Glu
1011
ATC TCT AAT ACT TTA TTG GAT ATA GCA ATG GAC CAM AAT TCT TTC CCC TCC AGT TTC I1. Sbr Amn Thr Lou Lou Asp Ile Ala MET Asp Gln Ann br Ph. Pro Ser SIr Ph.
1068
AAA TCA AAA TAT TTG AAT AGC AAT AAG ACA TCT TTA GCA Lye ber Lye Tyr Lou Asn bor Ann Lye Thr Sor Leu Ala BcoRI TTA CAA AAC TAT TGT ATT CAA TTG AAA CTG GAA GM TTC Iou Gln Asn Tyr Cys Il1 Gln Lou Lye Leu Glu Glu Ph.
AAA TCT CTC TTA GAC GCA Lys Sbr Leu Lau Asp Ala
1125
TAC CGT TCA CAA GAA TTG Tyr Arq br Gln Glu Lou
1182
GAA ACC ATG TAC AAT ACT ATC TTT GCT CAG TCC TTT GAC AGC GAT TCA TTG ACT TGT Glu Thr MET Tyr Ann Thr Ile Ph. Ala Gln S*r Ph. Asp br Alp Sbr Lou Thr Cys
1239
GTT TAC TCA AAT GCT ACT ACT CCA AAG AGC GCT ACG GTT TCA TCT GCG GCC ACA GAC Val Tyr Ser Asn Ala Thr Thr Pro Lys Sbr Ala Thr Val Snr SUr Ala Ala Thr Alp
1296
TAT TTC TCG GAT CAC ACT CAT TTA AGA AGG TTG ACC AAA GAT AOC ATT TCT CCA CCA Tyr Phe Sor Asp His Thr His Leu Arg Arq LAu Thr Lys Asp Sor le Sbr Pro Pro
1353
TTT GCC TTC ACT CCA ACC TCA TCT TCA TCC TCT CCA TCT CCA TTC AAT TCC CCT TAC Pro Thr Ser Sr Sbr Sor Slr Pro ler Pro Ph5 Ann S*r Pro Tyr
1410
AAG ACT TCA AGT TCA ATG ACG ACC CCA GAC TCT GCA TCA CAC CAT TCA CAT TCA OGT Lys Thr s0r Sor Sor MET Thr TShr Pro Alp Ser Ala Slr lis His Ser His Sor Gly
1467
TCG TTC TCT TCT ACC CAA AMT TCT TTT AAA AGG TCA CTG AOC ATC CCA CAA AAT Slr Phe Ser Sbr Thr Gln Ann Slr Ph5 Lys Arg Slr Lou Ser Ile Pro Gln Asn
OCA Slr
1524
AGC ATC TTT TGG CCA AGC CCA CTA ACT CCC ACC ACC CCA TCT CTA ATG TCA AAT AMA Sor Ile Phb Trp Pro Sor Pro Leu Thr Pro Thr Thr Pro ler Lou MT ler Ann Arg
1581
Ncol AAA TTA TTA CAA AAT TTA TCT GTG CGT TCA AAA AGA TTA TOT CCT GTT AGA CCC ATG Lys LAu Leu Gln Asn Lou S*r Val Ar; Slr Lys Ar; LAu Ph5 Pro Val Ar; Pro MET
1638
CCTCTOCCAC ATTTCCATAT TTGGCCTTGG GTTTTTGCCC TCATCTTTTT TTTTTCTTCC TCCTACCTAT -441 Xhol TTATAATTTG TATATCTSTA CTTTCCTCGA GCTTTTAATC TTCTTCTTAA ACATTTTCTT GTGTAGTATA -371 CTTTCACACA ATTTCTTTCT TGATTTTTTC TTCTACTACT GAGTCTGCCA GTCAAATGGA TTTCTGACGA -301 AAGACGACTA TACCCATTAG GAAACGAATT GCCCGAGTAG TCTCCTCTGC CGACTTAAAC CAACCTTTTT
-231
CTATTTCTCT TTTCTTTTCT CCCTCTTTTT TCTCTGTACT AGCATCCAAM AOCAAGCATC CATCCGAGTC
-161
CCAOTCGCAA TCTCACATCT CCAATTTAAC OTATCCATTG CATTTCCTCA TTCGGTTTAA
CTCCTCTGCA
-91
TTTCTTTOCT GACCCATAOC ATTTCTTACA TTCCATTGCA TCTCCCTTTT ACTCTCGTTC AAGACACTGA
-21
TTTGATACGCTTTCTGTACGATG GCC ATA TTO AAG GAT ACC ATA ATT AGA TAC GCT AAT GCA OlT Ala Ile LeU Ly Alp TShr Ile Il Ar; Tyr Ala Aen Ala
42
AGG TAT GCT ACC GCT AOT GOC ACT TCC ACC OCC ACT GCC GCC TCT CTC AGC OCT GCC Arg Tyr Ala TShr Ala Sbr Gly Thr Sbr Thr Ala Thr Ala Ala Sor Val Ser Ala Ala
99
TCA TGT CCT AAT TTG CCC TTG CTC TTG CAA AAG AGO CGG GCC ATT GCT AGT OCA AAG Ser Cys Pro ASn Lou Pro Lou LeU Lou Gln Lys Ar; Ar; Ala I1e Ala S*r Ala Lys
156
TCT AAA AAC CCT AAT CTC GTT AAA AGA GAA OTG CAA GCA CAT CAC TCA GCG ATC AGC Sr Lys Ann Pro Ann LOU Val Lys Arg 0lu Leu Gln Ala His eis S r Ala Ile sbr
213
GAA TAC AAT AAT OAT CAM TTG GAC CAC TAT TTC CGT CTT TCC CAC ACA GMA AGG CCG Glu Tyr Ann Ann Alp Gln LAu Alp Hni Tyr Ph, Ar; Lou Sr His Thr Glu Arq Pro
270
CTG TAC AAC CTG ACT AAC TOC AAC TCT CAG CCA CAA GTT AMT CCG AAG ATG CGT TTC Lou Tyr Ann Lou TShr Ann Ph- Ann Sbr Gln Pro Gln Val Asn Pro Lys MET Arg Pho
327
TTO ATC TOT GAC TOC ATC ATG TAC TGT CAC ACA AGA CTC AAT CTA TCG ACC TCG ACT Lou Ile Phe Alp Ph. Ile lET Tyr Cys Hls Thr Arg Lou Asn Leu Sr TShr s*r Thr
384
TTG TTC CTT ACT TTC ACT ATC TTG GAC AAG TAT TCC TCG CGG TTC ATT ATC AAG AGT Lou Ph0 LAu Thr Ph TShr I1e Lou Asp Lys Tyr Ser Sbr Arg Ph5 Ile Ile Lye Sor
441
TAC AAC TAC CAG CTC TTG TCC TTG ACC GCG CTT TOG ATT TCG TCC AAA TTT TGG GAC Tyr Ann Tyr Gln Lou Lou Ser Lou Thr Ala L,u Trp II* S*r Sor Lye Ph5 Trp Asp
498
T
Ph5 Ala Ph5 Thr
GCC ACT GCT CAC CCA TGC TCT GCC CCC ACC CM CTO AAA AAM AG TCA ACT TCC TCT Ala Thr Ala lis Pro Cys S*r Ala Pro Thr Gln Lou Lys Lys Arq ler Thr lSr o*r
1695
GTG GAT TGT GAT TTT MT GAT AGT AOC AAC CTC AM AM ACT COC TGA AMCACAAMA Val Alp Cys Asp Ph, Ann Alp Ser Slr Ann Leu Lys Lys Thr Arg BpaI MAAMTGCA TOTMCACA MTAMATTAA MTTTGCAM GCAMTACGT TAACATACAT TAATGATCTT
1753
TATACOTATA TTACATTTGC ATAGMATTAC
1893
TTTTTCTTTT TGTTTTCACA GGACTCATTA TTTCTTTAM
MMAAMAAAAAMTTATMA MCGCACAC
TCC AAG AAT AGA
ATO 0CC ACT TTG AAA GTC TTG CAA AAC TTG TGT TGC AAT CAA
eor Lys Ann Arg MET
TAT
TTTTTOCTCC TEACTTCCTC 1963
TAGTCACATT ATACAOTTTT STTCATCCTT CACAGACGTA TCTTAATACT MCTGTGCAT CMTATMTO 555
2033
GGTATAGCTT GMCATCTGC CCCTCTCTAT CTAATGTTTT TTTCTTGMAT TAOTTAGTAT TTGCTGTATA 2103
Ala Thr Lou Lys Val Lou Gln Asn Lou Cys Cys Asn Gln Tyr
CTAMTTTTAT AAM CAA TTC ACG Sbr Ie Lys Gln Pho Thr
ACT ATG GAA ATG CAT CTT TTC AAA TCA CTC GAT TGG TCC Thr MET 0lu MET Hls Lou Ph. Lye Sbr Lu Alp Trp Sor
612
ATC TGT CAG TCG CCA ACA TTC GAC TCC TAC ATC GAC ATC TTC TT TTC CAA TCT ACG I1. Cyn Gin Ser Ala Thr Ph. Asp Sr Tyr Ile Asp Ile Ph5 Lou Ph Gin Ser Thr lindIII TCC CCG TTA TCG CCT GOC GTT GTC CTT TCT GCC CCT TTG GMA GCT TTC ATT CAA CAG
669
TCT ATA
CTAMATACG ATTACTATCG
1823
TTATCMTAC AGACCAATGT TCMCAATTT
CTCATTGOTT TCCGCCMQCA TTCTSTAGC
2173
MCATTCCM TCTTCATCGC TACCTAAQC MCAMATCC CCATCOTCGT CCTOATATTT GATCTTGOTG 2243 ATTGGTGAM TGTTGTTATT
ATOTGTATTC MAAATTTTAG
TCCAMCTTT TTCTACCMA ATGTGAAGA TC/
MTTGATCGC CATTATCMG
OCOTCAAAAT
2313 2345
726
Ser Pro Lou ler Pro Gly Val Val Lou ler Ala Pro Leou Glu Ala Phe Ile Gln 0ln
AAA CTG 0CC TTA TTA AAT AAC GCT GCT GGT ACT GCT ATT MAT AAA TCG TCC TCT TCT Lys Lou Ala Lou Lou Asn Asn Ala Ala Gly Thr Ala Ile Ann Lys Slr ser ser ser
783
CAR GGC CCC TCT TTG AAC AC AAC GAM ATC AAA TTG GOT GCC ATT ATG TTG TGC GAG Gln Gly Pro Ser Ltu Asn Ie Asn Glu I1- Lye Lou Gly Ala Il. MET Lou Cyn Glu
840
TTA OCT TCC TTC AAT CTC GM TTA TCA TTT MA TAT GAT CGT TCA CTA ATT GCG CTG Lou Ala eor Ph5 Ann Leou lu Lou Ser Ph- Lys Tyr Asp Arg Ser Lou Ile Ala IAu
897
Fig. 3. The sequence of WHII +. The C to T transition of WHIJ-J (base 1210) is indicated. The 5' end of the sequence overlaps with the sequence deposited in EMBL by Dumont et al. (1987) for CYC3, and the 3' end overlaps with the sequence of CDC24 (Miyamoto et al., 1987). Possible sites for PRTF binding (consensus TTTCCTAATTAGGAAA-Bender and Sprague, 1987) are found between nucleotides -972 and -942; -683 and -669; -292 and -277; and -118 and - 103.
strain, producing transformants with one mutant and one wild-type copy of the region (WHIl + WHII-I strains). These transformants had a volume of 27 4 2 ,^tm3 (Table I), and a short GI (Figure 6), and were indistinguishable from WHII-I mutants. This shows the mutation is fully dominant. To show that the nonsense mutation in FUNIO was the WHII-I mutation, we removed the EcoRI -HpaI restriction fragment carrying the C to T transition from pBF30. Cells were transformed with the remainder of pBF30. Because gap repair would replace the missing sequences with wild-type sequences (Rothstein, 1983), the transformants recovered would again have two copies of the CYC3 and FUNIO regions, but this time without the nonsense codon. The transformants had an average volume of 35 i 3 /tm3, versus 42 /tm3 for their parental strain, and 27 i 2 ,tm3 for
4338
isogenic WHII + WHII-] transformants. Since strains with two doses of the wild-type gene (i.e. 2 x WHII + strains-see below) had volumes of -35 /.tm3, this result shows that the nonsense mutation is the WHII-I mutation.
WHI1 mRNA is not cell cycle regulated, but does respond to a factor Yeast cells increase in volume as they progress through the cell cycle. When an exponentially growing culture is fractionated on the basis of cell volume, each fraction represents a different part of the cycle. Centrifugal elutriation (Hereford et al., 1981) was used to separate growing cultures of MATala WHIJ +/WHII + diploids into 11 fractions. RNA was prepared from each fraction, and subjected to Northern analysis. Each fraction contained the same proportion of WHI1 mRNA (data not shown). Thus, the gene
Control of division by cell size Urchin cdc13 Clam A
mttrrltrqhllantlgnndenhpsnhiaraksslhssenslvngkkatvsstnVpKKrhaLddvsnFhN
WHIl Urchin cdc13 Clam A
mnlhgeskhTfNneNvsArLggKsiavQRpaqrAalgNisnvvRtaqaGskkvVkkdtrQkamtktkats
WHI1 Urchin cdc13 Clam A
yanaryatasqtstataasvsaascpnlplllqkrraiasakskNpNlVkrElqAHhsaiSEYnnDqLdh slhaVvglPvedLPtEMR tpvdaEdaieafSQqliaLQvEDIdkDDgdNPqLcSEYakDIYlY hgvdafhkdqAtIPkkLKkdvDErvvskDIpklhrdsvEspesrDwDDLiiEDwADPLMVSEYvvDIFeY ptshVtdiPaAlttlQrvpltEvpqspdiIsledsmeSpmildLpezEkplDreAviLtvpEYeeDIYnY
WHI1 Urchin
cdc13 Clam A
yfrLshterplynltNfnsQpqvnpIMRfLIfDfIMycHtRLNLstsTLFLtftILDKYsSrfiIksyny LRrLEVEnMvpanYLDRQetqiTg-RMRlILVDWLVQVHlRFHLLqETLFLtVqLIDRFLaeHSVSKGKL LneLEIZtMpsPtYMERkELaw--RMRgILtDWLIEVHsRFRLLpZTLFLaVNIIDRFLSlRvcSlnKL LRqaEMknrakPgYMkRQtDItTs--MRcILVDWLVEVseedKLhrETLFLgVNyIDRFLSkiSVlRGKL
WHIl Urchin cdc13 Clam A
QLVGVtAMFIASKYEEMYPPEInDFV--YITDNaYTkaQlrQMEZiaMLKgLkYkLgkPlcLHFLRRnSKA QLVGIAALFIASKYEZVmcPsVqNrV--YMaDggYdeEEILQaZRyILRVLeFNLAyPnpMNFLRRiSKA QLVGaAsMFLAaKYEEIYPPDVkEFa--YITDDtYTsQQVLrMEHlILKVLtFDVAvPttnwFcedflKs
WHI1 Urchin
cdc13 Clam A
WHIl Urchin cdc13 Clam A
MalgtrnmN
msqpfalhhdgenqMqRRgkmntrsngLsg mailkdtiir
kegvpLaskntNVrHttAsVstRrAleEJCSiipAtDDepasXCKRrqEsvfnsSV Slpqltsss qkraaLgviTnqvnQqvriqpsRaAkpksSefniqDENaftKxnaktfGqqpSqfSvfvDptpaapvqka
QLLsLtALwIsSKFwDsknrmatlkVlqnLccNqYsikQfttMEmhLFKsLdwsIcqsatFDsyidiflf
qstspl
pvlalaIQkakdSiAlgAInLiKlsLnyynsnlWe-NInlaleencqDLdi dfyDiQtRTVAKYLVEIgLLDhkL-LpYpPSQqcAAAMYLAReYCsrgP ---WnrNLVHYSGYEEyQLis
agvDaQkHTLAKY1MEITLpEysM-VqYsPSEiAAAAIYLsmtLLdpEthssWcpkMtHYSmYsEdHLrp
cdaDdklKsLtmFLtELTLIDmdayLkYlPSitAAAALcLARysLGiEP ---WpqNLVKktGYEigHFvd
kLseIsntLLdiamdQnsfpssfksKYlnsnktslakslldalqnyciqlkleeFyRsqeletmyntiFa
IVqKIVqiLLRdDSasQky-sAVktKYgSsK ----------------------- 4MRiSgiaqLdsslLk VVKKMInyLqKpvqHE -----AffkKYaSkK----------------------- FMKaSLfvrdwIkkns cLKdLhktsLgaESHQQq ---AVqeKYkqdKyhqvsd ----------------- FsKnpVphnLaLlaL STOP
WHIl Urchin
qIsfdsdsltcvysnattpksatvssaatdvFsdhtHLrRltKDsisppfaf tPtssssspspfnspykts
cdc13
qiaqgsnE iplgddaD ------------------ edvtFhkakRIaHdmKDeew
WHIl
ssmttpdsashhshsqsfsstqnsfkrslsipqnssifwpspltpttpslmsnrkllqnlsvrskrlfpv
WHIl
rpmatahpcsaptqlkkrstssvdcdfndssnlkktr
Fig. 4. Alignment of WHII with three cyclins. The complete sequences of WHII. sea urchin cyclin, clam cyclin A and cdc13 are shown. Upper letters represent conserved residues; bold upper case letters represent highly conserved residues, as defined in Materials and methods. Regions enriched in P, E, S, T and/or D (PEST regions) are underlined. Basic residues associated with the PEST regions are italicized. Also italicized is the WHII sequence nktslakslld. which is similar to a sequence found 95 residues more N-terminal in the other cyclins. The position of the STOP codon in WHIJ-J is shown. Further details are given in Materials and methods. case
is constitutively transcribed throughout the cycle. To see if factor could modulate WHI] expression (an experiment suggested by the a factor resistance of WHII-], see below), we treated MATa WHI1+ and MATa WHII-I factor for 1 h. RNA was cells with 5 x 10-6 M extracted and examined by Northern analysis (Figure 5). It appeared that the WHII mRNA was induced 2- to 3-fold. The amount of transcript seemed to be slightly lower in the mutant than in the wild-type. Perhaps the nonsense mutation, which would prevent translation of the last third of the message, reduces mRNA stability. In view of results presented below, it is important to note that the WHII-I gene certainly does not produce significantly more mRNA than the wild-type gene. a
a
The region upstream of WHI] was examined for various consensus sequences associated with ax factor induction. There were several six out of eight matches to ATGAAACA or its complement (Trueheart et al., 1987; Dietzel and
Kurjan, 1987), but no exact copies. There were four regions with some similarity to P boxes (Bender and Sprague, 1987). A P box is thought to be the binding site for the Pheromone Recognition Transcription Factor (PRTF), which, despite its name, may be a general transcription factor (Bender and Sprague, 1987). WH1 + is a dose-dependent activator of Start In the course of other studies, B.Gallay and R.Rothstein (personal communication) constructed plasmid pWJ3 10 4339
R.Nash et al. A
Table I. Effect of WHII alleles on cell volume and i.:
.- .-X
Genotype
whil-310(A) WHII + 2 x WHIl + WHII-I WHII + WHII-J 2 x WHII-I
kb .:
.~~~~~~~~~~~~~~~~~~~~~~~A _-
Fig. 5. The WHII transcript is a factor inducible. 5 Ag of total nucleic acid from a wild-type (W.T.) or WHIi-i strain either non-induced (-) or induced (+) for 1 h with 5 x 10-6 M a factor was electrophoresed on a 1% agarose gel, and transferred to nitrocellulose. The filter was probed with a cloned WHII gene. The 2.1-kb WHII transcript is indicated. The markers (lane M) are single-stranded DNA. Chromosomal DNA is indicated (C). WHII-I cells, being small, have a higher DNA to RNA ratio than wild-types. After autoradiography, the probe was removed, and the filter was re-probed with cloned ADHI DNA as a loading control.
(Figure 2), and deleted FUNJOin a diploid. Upon tetrad analysis, they found that FUNIO was not an essential gene. We repeated these experiments (Materials and methods) and found that a whil deletion increased cell volume by - 25 %. This suggested (i) that the WHIl protein was an activator of Start, since its absence delayed Start, making cells larger; and (ii) that the mutant WHIl-1 protein might be an overactive protein, since it had the opposite effect of a deletion. To confirm and extend these conclusions, we surveyed the effects of over- and under-expressing the wild-type and mutant genes. This was done by integrating tandem extra copies of the genes at the chromosomal WHII + locus (Materials and methods). Effects of varying WHI] + or WHII-I expression were assayed in two ways. First, cell volumes were measured. Second, cells with DNA stained with propidium iodide were analyzed by flow cytometry to measure the lengths of GI and G2 phases. Data from many experiments are summarized in Table I, and particular experiments are shown in Figure 6. Gratifyingly, the length of GI was well correlated with cell volume, and WHII-I did appear to act by accelerating Start, thus shortening the time cells spent in G1. The results can be summarized as follows: (i) cells bearing a whil-310
4340
Cell volume
GI
phase % of cycle
(Am3)
Normalized volume
spent in G
55 t 5 44 ± 3 35 3 27 ± 2 27 25
1.2 1.0 0.80 0.61 0.61 0.57
37 27 NT 19 15 0 to 10
NT, not tested. 4 represents the range over at least eight measurements. Where is not indicated, less than eight measurements were done. For most constructions, each transformant behaved similarly, but for 2 x WHII-I strains, some transformants repeatably had a GI of -0, while others had a GI of - 10%.
deletion (the allele generated by deletion/disruption with plasmid pWJ310; Materials and methods) were -25% larger than wild-type cells, and had the longest GIs; (ii) 2 x WHII + cells had a volume -20 % smaller than wild-type cells; i.e. a partial Whi- phenotype; (iii) WHII-I cells were very small, 40% less than wild-types, and had a short G1; and (iv) at least some 2 x WHII-I cells had no visible G1 phase, but were only marginally smaller than WHII-I cells. Our results with 2 x WHII-I transformants were variable. Some showed no G1 phase, while others from the same transformation spent -10% of their cycle in G1. Only one 3 x and one 4 x WHII-I strain were obtained; these had very short but visible G1 phases (data not shown). The site of plasmid integration was the XhoI site in the WHII + promoter; small alterations in the promoter may have caused variations in WHII-J expression. Variations in culture conditions may also have contributed to variability. Both alleles were also cloned into the high copy number vector YEp352 (Hill et al., 1986). However, the vectors were mitotically unstable, and culture conditions had to be modified to maintain the plasmids. Because of these difficulties, the results obtained with YEp352 were not directly comparable to other results. As best we could judge, 5-10 doses of WHI] + resulted in a phenotype more extreme than the 2 x WHII + construction, but slightly less extreme than a WHII-I mutant (data not shown). Five to ten doses of WHII-I shortened G1, but did not reduce size below 27 jtm3. All of these results are consistent with the idea that the wild-type WHII + gene is a dose-dependent activator of Start, and that the mutant WHII-I gene has an over-active but qualitatively similar function. The manipulations of WHII + and WHII-I had surprisingly. little effect on culture doubling times. Wild-type cultures typically doubled in -90 min, while isogenic WHII-I or 2 x WHII-I strains doubled in 92-94 min. Thus, cells lacking any observable GI were nevertheless healthy. The time lost in GI was simply spent in some other portion of the cell cycle. We have preliminary cytological evidence that the part of G2 immediately before nuclear division was elongated in WHII-I cells. Deleting the gene did have a small effect; whil-310 strains had doubling times of - 100 min, versus 90 min for their isogenic parents, but this was a small change compared to the effect on size and G1. Thus, GI can be expanded or -
Control of division
0
X -Jc4or
Molarity
'Zx
3x -6
-
-6
whi 3a
... -1
,
;Ii
"r
2
3x -5
-5
r
WT
L
by cell size
aEnE
WHII1-1a c n n x WHI 1-I aE
sw 7, WHI 1-1 a" h
,:2 i/
0
%k X .r--
_,,' -/ _ ~
oC; Factor Molarity O 0 9 -8 -7 -6 -5
sst2flaEEEEE h 3 sst2I2EHEM
-J-I sst2 Fig. 7. a Factor resistance of mutants. Part A: concentrations of a factor, from left to right, are 0, 10-7 M, 3 x 10-7 M , 10-6 M, 3 x 10-6 M, 10-5 M, 3 x 10-5 M. Cells grew from an original inoculum of 5 x 103 cells (Materials and methods). Awhil is the
1~~~~~~~~~~~
whil-310 deletion strain. Another whi mutant, whi3, is included for
Fig. 6. Flow cytometry of mutant cells. Cells were stained with propidium iodide (Materials and methods), and the fluorescence per cell was measured. The axis is the number of cells; the x axis is the intensity of fluorescence. Peaks due to cells in GI and G2 phase are indicated. The G2 peak actually includes cells in G2, M and cytokinesis. The pairs of panels a and b, c and d, and e and f show parental strains and isogenic transformants. Peaks drift to the left as cells get smaller; this is an artefact due to a small amount of fluorescence from cytoplasmic cell wall material.
contracted over a wide range (0-40% of the cell cycle) without otherwise greatly affecting the cells. WHI1-1 cells
are
resistant to
a
factor arrest
WHII-] x WHII-I crosses often produced rather few diploids. Since WHII-] seemed to activate Start, and since factor (a yeast mating pheromone) normally inhibits Start, we wondered whether WHII-I mutants might react aberrantly to ae factor. We challenged MATa WHII-I cells with factor, and assayed several responses, including cell cycle arrest, mating ability and induction of the a factorinducible mRNAs for SS72 (Dietzel and Kurjan, 1987) and sigma (Van Arsdell et al., 1987). The most striking result was that strains carrying WHIJ-J could not be permanently arrested by ca factor. While wild-type cells were permanently arrested by -3 x 10-6 M a factor, WHIJ-J cells efficiently formed colonies on plates containing 3 x 10-5 molar factor (Figure 7). a
a
a
comparison. Part B: concentrations of at factor, from left to right, are 0, 10-10 M, 10-9 M, 10-8 M, l0-7 M, 10-6 M and lo-5 M. A few clonal colonies are visible in the sst2 and whi3 sst2 rows. These uncharacterized resistant mutants not present in other experiments. were
Despite the lack of permanent arrest, examination by microscopy showed that when a MATa WHII-I culture was first exposed to ax factor the percentage of budded cells dropped, and some cells took on the characteristic shmoo shape of cycle-arrested cells. Growth curves showed a pause in division in response to pheromone, and analysis with the Channelyzer showed a large increase in cell volume shortly after application of pheromone. A GI-less 2 x WHII-I strain had a similar phenotype. For these reasons, we believe that at least some WHII-I cells probably arrest when a factor is applied, but then recover quickly and continue cycling. While WHII-I cells were clearly defective in cell cycle arrest, both MATa and MATa WHII-I cells mated reasonably well (- 5-fold worse than wild-type). When treated with 5 x 10-6 M a factor, MATa WHIJ-l cells induced the SS72 (Figure 8) and sigma (data not shown) transcripts to the same levels as did wild-type cells. Both MATa and MATa WHII-I cells induced shmooing in strains of the opposite mating type. Finally, MATa/a WHIJ-JI WHI]-l cells sporulated, albeit with a reduced frequency attributable to their small size (Calvert and Dawes, 1984). Thus, the a factor resistance phenotype seemed to be 4341
R.Nash et al.
Table II. Ability of WHIl-1 cells to enter a heat-resistant phase
Genotype WHII + WHII-I
% Viability Exponential cells
Stationary cells
0.5 2.5
91 96
Cells were grown to 3 x 107 cells/ml and then heat-treated (exponential phase), or to 2 x 108 cells/ml, then incubated a further 48 h, and then heat-treated (stationary phase). For heat treatment, cells were incubated at 45°C for 0 min or for 40 min, and then cells were spread on YEPD plates. The number of survivors after 40 min was compared to the 0 min control.
A number of ts lethal cell division cycle (cdc) mutants are known that cannot complete Start at the restrictive temperature (Reed, 1980). Double mutants between WHI]-l
and five of the cdc Start genes-cdc28-17, cdc36-16, cdc39-1, cdc61-1 and cdc63-1-failed to grow at 37°C, showing that Start was still part of the WHI]-l cell's cycle. 2 x WHI1-1 cells regain G1 at low growth rates When growth is slow, cells adapt by lengthening GI (Hartwell and Unger, 1977; Johnston et al., 1977; Carter and Jagadish, 1978). It was of interest to see how GI-less, 2 x WHII-] cells would adapt to slow growth. When the mass doubling time was increased to 260 min by using glycerol as the carbon source, 2 x WHII-I cultures did contain a majority of unbudded cells, i.e. cells apparently in GI. At all growth rates, however, the proportion of unbudded cells was less in 2 x WHII-I or WHII-I strains than in wild-type strains (data not shown). This result suggests that even 2 x WHII-I strains have a critical size for Start, but that this critical size is less than the birth size in fast-growing cultures. The critical size is greater than birth size in slow-growing cultures.
Discussion
Fig. 8. WHIJ-J strains respond to a factor induction. As Figure 4, except that the filter membrane was Nytran. The SS72 transcript and chromosomal DNA (C) are indicated. The SS72 transcript appears to be a doublet; this is probably a gel artefact due to the proximity of a large amount of rRNA.
specifically due to a defect in cell cycle arrest. Other a factor responses seemed normal, as did other mating functions. The fact that a factor could make WHIl-l cells pause but not stop suggested that WHI]-l cells might recover from the effects of a factor very quickly. To address this, we examined MATa WHII-I sst2 double mutants. sst2 cells are supersensitive to a factor, and may be defective for recovery (Chan and Otte, 1982a,b). Several MATa WHII-I sst2 strains were constructed (Materials and methods) and challenged with at factor. Like their parental MATa WHII-I SST2 strain, they were entirely resistant to permanent arrest even by 10-5 M a factor (Figure 7). Although WHII influences the a factor response, it is not a haploid or mating type-specific function. MATa, MATa and MATa/ae cells all show the WHII-I phenotype, and the WHII transcript is present in all three cell types (data not shown). WHI 1-1 cells can enter stationary phase, and require the CDC Start genes Since WHII-I cells were defective in a factor arrest, they might also be defective in other kinds of cell cycle arrest. By the criterion of heat shock resistance, WHII-I cells arrested normally in stationary phase when grown to saturation in YEPD (Table II).
4342
Wild-type cells cannot bud until they have reached a critical size of -44 utm3. The WHII-I mutation reduces this critical size to -27 ,Im3, and shortens GI drastically, but does not affect culture doubling time. The WHII-I allele encodes a truncated version of the wild-type protein. Since complete deletion of the gene has a phenotypic effect opposite to that of the WHII-I mutation, and since over-expression of the wild-type gene has an effect similar to that of WHII-], we believe that the WHI1-l protein is a hyperactive or longlived version of the wild-type protein with a qualitatively similar activity. The fact that cell volume and the length of GI are proportional to the dose of WHII + (for doses between 0 and 2) is consistent with the idea that the concentration of WHII might be the metric by which commitment is determined. One of our most surprising findings was that two doses of WHII-I could eliminate any visible GI in fast-growing cells. The GI-less cells were healthy, with apparently well coordinated cell cycles. Start has no apparent size requirement in these strains, i.e. what we had taken to be the major cell cycle control event was entirely relaxed, and yet the cells were not greatly inconvenienced, and still efficiently coordinated growth with division. This must mean that there are multiple, redundant controls preventing overfrequent cell divisions. Perhaps Start cannot occur until after cytokinesis, and this prevents overlapping cell cycles (budded buds). Also, we have preliminary cytological evidence for a control at the beginning of nuclear division. In S.pombe, the major cycle control point is at the beginning of nuclear division; a cryptic GI control is seen only under conditions of slow growth (Nurse and Thuriaux, 1977), or when the G2/M control is relaxed by the wee] - mutation (Russell and Nurse, 1987). It may be that S.pombe and S. cerevisiae have controls at both points, but differ in which control is the most restrictive in fast-growing cells. Singer and Johnston (1981) have previously argued that G1 is dispensable in S.cerevisiae, and some animal cells lacking a GI phase have been found (Robbins and Scharff, 1967; Liskay, 1977).
Control of division by cell size
Deleting WHII + is not lethal. The phenotype of deletion strains argues that WHI + is an activator of Start, but the fact that the cells are alive proves it is not an essential activator. Again, the most likely explanation is that there are multiple, redundant controls for activating Start. Perhaps S. cerevisiae, like the surf clam (Evans et al., 1983), has two cyclins, and either of them is sufficient for activation of division. We are using genetic and molecular methods to find the cell cycle controls that back up WHI1 + when it is either missing or over-active. One other Whi mutant has been characterized. The whi2 mutation (Sudbery et al., 1980; Kelly et al., 1988) makes cells small when they grow on poor carbon sources, but not when they grow on glucose medium. The mutation prevents cells from taking on stationary phase characteristics, such as heat shock resistance (Kelly et al., 1988). The sequence of the gene shows no similarity to that of WHI] (Kelly et al., 1988). WHII-] whi2 double mutants are viable (Sudbery et al., 1980). Sequence analysis suggests that WHIl is a cyclin. Cyclins were originally identified in clam and sea urchin embryos as proteins which changed in abundance as the cell cycle progressed (Evans et al., 1983). The proteins are synthesized continuously, accumulate during S and early M phase, and then are catastrophically degraded near the end of M phase (Evans et al., 1983). Clam cyclin A (Swenson et al., 1986) and sea urchin cyclin (Pines and Hunt, 1987) have been cloned and sequenced, and mRNAs have been produced in vitro. When these mRNAs are injected into quiescent Xenopus oocytes (which are arrested in the first meiotic prophase), they push the oocytes through meiosis I. Thus, these cyclins are direct activators of meiosis, and probably mitosis (Swenson et al., 1986; Pines and Hunt, 1987). S.pombe cdc]3 - mutants are blocked at the G2/M boundary at the restrictive temperature (Booher and Beach, 1987, 1988), and this is consistent with the idea that the cdc]3 + cyclin is an activator of mitosis. We have taken the WHIl protein to be an activator of Start, a GI commitment point, while the other cyclins seem to activate an event at the G2/M boundary. This discrepancy has several possible explanations. First, perhaps there are several classes of cyclins, and some of them act in G1. Second, it has been argued that S. cerevisiae has an extended mitosis that begins at about the same time as S phase (Nurse, 1985). Perhaps the mitotic event controlled by cyclins occurs after S phase in most cell types, but at Start or shortly afterwards in S. cerevisiae. A third possible explanation has to do with the properties of the cdc28 mutation (see below). CDC28 is one of the most important proteins required for Start in S. cerevisiae. CDC28 has protein kinase activity when it is in a complex with several other proteins (Reed et al., 1985; Mendenhall et al., 1987). We have not examined the relationship between WHII and CDC28, except to say that WHIll- cannot completely suppress a cdc28-17 mutation. In particular, we have not addressed the possibility of partial or allele-specific suppression. However, an interaction between WHIll and CDC28 is to be expected for several reasons. (i) In S.pombe, some alleles of cdc13 - can suppress some alleles of cdc2 - (Booher and Beach, 1987) (cdc2 is a CDC28 homolog-Hindley and Phear, 1984). (ii) cdc2 + in high copy number can suppress some alleles of cdc]3 (Booher and Beach, 1987). (iii) Cyclins are thought
to activate Xenopus maturation promoting factor (MPF), (Swenson et al., 1986; Pines and Hunt, 1987) and MPF is a complex containing the Xenopus homolog of the cdc2/ CDC28 protein kinase (Dunphy et al., 1988; Gautier et al., 1988). Since the other cyclins interact with or activate their cognate CDC28 homolog, WHIl may turn out to be an activator of CDC28 kinase activity. In this regard, it is
interesting to note that Mendenhall et al. (1987) have postulated an 'exchangeable factor' needed for activation of the CDC28 kinase complex. This factor was apparently available only at some cell cycle stages, and so had some of the attributes of a cyclin. CDC28 is required primarily for Start (Hartwell et al., 1973; but see also Piggott et al., 1982 for an alternative view), while cdc2 is required for both Start and for mitosis (Nurse and Bissett, 1981), and MPF is only known to be required for mitosis. Therefore, if the function of a cyclin is to activate the CDC28/cdc2/MPF protein kinase, then the apparent time of action of the cyclin would be the time at which the kinase is required. Thus, in S. cerevisiae WHI1 would act at Start because that is where CDC28 is required, while in Xenopus, injected cyclins act at G2/M because that is where MPF is required. This fails to explain why S.pombe cdc]3- mutants do not have a Start defect. How might WHII work? A speculative model is that it is synthesized continuously, and accumulates during G1. While accumulating, it might also be modified by glycoslyation, phosphorylation, or limited proteolysis, and these modifications might increase the protein's activity. The rate of synthesis or modification might be correlated with cell size. When a sufficient amount of active WHII had accumulated, the cell would be pushed through Start. At some later point, the active protein would be destroyed, as other cyclins are, and the cycle of events could start again. The PEST sequences found in WHI1 and the other cyclins might be the signals for destruction; however, the idea that PEST regions are proteolytic signals has not been tested experimentally. The effect of the WHII-I mutation can be explained in the following way: the wild-type, PEST-ridden C terminus may either inhibit the wild-type protein's activity, or shorten the protein's half life. The WHIl-l protein, which lacks the PEST tail, would therefore have a greater activity or longer half life, so that WHII-I mutant cells would always have a relative excess of WHI I -I activity. This would tend to push cells through Start prematurely, or even constitutively, as in the 2 x WHII-I strains. There are several reasons why WHIl-] strains might be resistant to a factor arrest. First is the possibility that the reduced cell surface area reduces the number of a factor receptors. This is unlikely for many reasons; e.g. (i) while the absolute surface area is small for a WHIl-] cell, the ratio of area to volume is relatively large; (ii) receptor number is not necessarily limited by area; (iii) enough signal gets through to induce other mating responses; (iv) fractional occupancy of available receptors is sufficient to arrest sst2 cells (at 10-9 M ca factor), but nearly full occupancy (3 x 10-5 M) does not arrest sst2 WHIl-] cells (the dissociation constant for a factor is 6 x 10-9, Jenness et al., 1986). A second possibility is that cells are only sensitive to a factor arrest during a short window in GI, and since WHII-I cells have little or no GI, they cannot be arrested (the Window of Vulnerability hypothesis). An advantage of -
4343
R.Nash et al.
this model is that it correctly predicts the behavior of whil-301, which has a long G1, and is relatively sensitive to a factor (Figure 7). A third possibility, which we favor, is that WHII-] and a factor are antagonists. That is, WHII-I is an activator of Start, and at factor is an inhibitor. The relative amounts of these effectors determine whether Start occurs or not. In the WHIJ-J mutant, the activator is so powerful or abundant or long-lived that a factor is incapable of permanently inhibiting Start. We believe that this is why WHII-I suppresses sst2. Two observations hint that the wild-type gene may assist in normal at factor recovery. First, WHII + deletion strains are - 3-fold more sensitive to a factor than wild-type strains. Second, there is some evidence that the WHII + transcript is induced -2-fold by a factor. Induction of the WHII + activator would help overcome oa factor-mediated inhibition of Start. The WHI] + gene has many of the properties expected of a controller of cell division. Its various alleles change the volume at which commitment to division can occur, and commitment is sensitive to the concentration of the gene product. While many other genes have been identified that are clearly involved in commitment to division, to our knowledge this is the only gene identified that clearly affects the time of commitment, i.e. WHII + is a regulator of commitment, not just a necessary part of the commitment machinery. With our present molecular and genetic knowledge of WHII +, there are several routes that can be taken to find molecules and genes that interact with it; we hope that such approaches will lead to an understanding of the molecular nature of the commitment event.
Materials and methods
the transformants to a wild-type strain. After sporulation, Whi+ Sst- spore clones were recovered at the expected frequency. Start cdc mutant strains were obtained from R.Singer. Strain DK17-4b (cdc24) was obtained from D.Kaback.
Plasmids; cloning of WHI1+ and WHI1- 1 DNA Plasmid pWJ310 was obtained from B.Gallay and R.Rothstein. They constructed it by cloning an AvaI-BglII fragment of chromosome I into a modified (filled-in Eco RI site) pUC 18. The SalI-EcoRI chromosomal fragment was then removed, and replaced with a Xho I-linkered URA3 fragment. We cloned WHII-I and WHII + by integrating pWJ310 at the chromosomal HpaI site in the WHI locus. Total DNA was prepared from these transformants, digested with Apa I and ligated at low concentrations. The ligated DNA was used to transform Escherichia coli to ampicillin resistance. A plasmid consisting of pUC18 with a 5.5-kb insert of chromosome I DNA but no URA3 gene was recovered. Of the 5.5-kb insert, some was from pWJ310, but the central 4-kb HpaI-ApaI portion was derived from the transformed yeast strain. This was the source of DNA for most experiments, including sequence analysis. Plasmid pBF30 was constructed by choosing one of the WHII-I pUC18 clones described above, and cloning a Xhol-linkered URA3 gene into the SalI site in the insert. The wild-type WHII + gene was also cloned from the lambda phage Cla (Coleman et al., 1986) provided by D.Kaback. A 7-kb BglII fragment carrying WHII + was cloned into the integrating vector YIp352 (Hill et al., 1986); this was the gene used for most but not all of the 2 x WHII + constructions.
Preparation of DNA The alkaline lysis method (Maniatis et al., 1982) was used to prepare plasmids from E. coli. Yeast DNA was prepared as described by Holm et al. (1986). Southern and Northern analysis Southern and Northern analysis was by standard procedures (Maniatis et al., 1982). The number of plasmid copies integrated at the WHII locus was assayed by measuring the length of the WHI I-hybridizing Bgl II band after Southern analysis. BglII gives a WHII chromosomal fragment of 7 kb, but does not cut within the gene or within the integrating plasmids. Because a 4 x WHII-I strain produces a BglII band of 34 kb, the gels used for analyzing potential 3 x and 4 x strains were composed of 0.5 % agarose, and were run for 60 h at 1 V/cm. Resolution up to 50 kb was achieved. -
Media The media used have been described (Futcher and Carbon, 1986). YEPD buffered at pH 4.5 with 30 mM sodium succinate was used for mating assays and pheromone response tests.
Strains Strain S673a (MATa WHill-i Iys2) from P.Sudbery was crossed to strain LL20 (MAToh leu2 his3 canl) from G. Fink. A segregant was back-crossed to LL20 three times to generate the BF328 tetrads, and four times to generate BF334 tetrads. A his4 and a ura3-52 marker were crossed and four times back-crossed into the BF334 background from strain BWG1-7a (MATa adel ura3-52 leu2 his4) obtained from J.Boeke. All strains used for size measurements were at least 15/16ths isogenic with strain LL20. The circular integration and transplacement techniques described by Rothstein (1983) were used to construct strains with alterations at the WHII locus. Transformations were by the lithium acetate method of Ito et al. (1983). WHII + WHII-I strains were constructed by cutting the integrating plasmid pBF30 with XhoI, transforming a WHII + ura3 strain, and selecting for Ura+. Southern analysis confirmed that most transformants had two tandem copies of the CYC3-WHII region. whil-310 deletions were made by cutting pWJ3 10 with HpaI and Sph I, transforming WHII + ura3 strains, and selecting for Ura+. Southern analysis confirmed that the chromosomal region normally found between the Sal I and the Eco RI sites was missing, and had been replaced by URA3. 2 x WHII + strains were constructed by cutting plasmid YIp352-WHII + (generated by cloning a BglII chromosomal fragment carrying WHI1 + into YIp352, Hill et al., 1986) with XhoI, and transforming as above. 2, 3 and 4 x WHII-I strains were constructed by cutting pBF30 with XhoI, and transforming WHII-I strains. The number of integrated genes was measured by Southern analysis (see below). Size comparisons were then made between transformants and their parents, and so comparisons were between truly isogenic strains. WHII-1 sst2 double mutants were constructed by cutting the plasmid pBC33 (obtained from W.Courchesne and J.Thorner) with the enzyme SstH, which released a disrupting fragment marked with URA3, and then transforming various MATa WHII-i strains. To confirm that transformants were indeed sst2 disruptants, we did Southern analysis, and also crossed
4344
Northern analysis was done on total nucleic acid. Total nucleic acid was extracted by vortexing cells with glass beads in the presence of phenol. RNA was measured by the fluorometric method of Morgan et al. (1979) with the modification that fluorescence was read both before and then after addition of RNase A to the assay tube so that equal quantities of RNA could be loaded in each gel lane. Sequence analysis WHII-] and WHII + DNA were cloned into the single strand producing
plasmids pUCl 18 and pUCl 19 (Vieira and Messing, 1987). Nested deletions were made by the method of Henikoff (1984). Single-stranded DNA was recovered with the use of the helper phage M13K07 (Vieira and Messing, 1987). Both strands of both clones were sequenced by the dideoxy method (Sanger et al., 1977) with a Sequenase kit (US Biochemicals).
Computer analysis and protein alignments Pairwise alignments between WHI 1 and each of the cyclins were done using the IALIGN program of the Protein Identification Resource at Georgetown University, Washington, DC. These pairwise alignments were adjusted by eye to produce the best four-way fit. Conserved amino acid replacements were defined using the mutation data matrix by Dayhoff et al. (1979). The matrix gives scores ranging from -8 (W:C) to 17 (W:W) for each possible amino acid match or mis-match. The scores are based on the frequency of actual amino acid replacements between present day proteins and proteins inferred as common ancestors. In Figure 4, amino acid pairs with a score of 2 or 3 are shown in upper case, and pairs with a score of 4 or more are shown in bold upper case. Identical matches have scores between 2 and 17. Non-identities with a score of 2 or 3 are the following: R:H, R:K, R:W N:D, N:H, D:Q, D:E, Q:E, Q:H, I:L, I:M, L:F, L:V and M:V. Non-identities with a score of 4 or more are the following: I:V, L:Mand F:Y. Figure 4 was constructed by establishing a consensus sequence for the three cyclins. If two of the three had identical or conserved (conserved meaning a mutation data matrix score of two or more) amino acids at the same position, that was considered a consensus amino acid. WHIl was then aligned with this consensus, and WHII amino acids were written in lower
Control of division by cell size case, upper case, or bold upper case, as appropriate to the particular match between WHI I and the cyclin consensus. PEST regions were found by assigning P, E, S and T a score of 2 each, D a score of 1, and all other amino acids a score of 0. The proteins were scanned with a window 15 amino acids wide. All windows with a score of 16 or more were noted and the extent of the PEST region was deternined by eye, maintaining at least 50% PEST residues. Associated basic residues were found by eye. Figure 4 does not show several weak PEST regions not tightly bounded by basic residues. Note that some PEST regions are highly enriched in just one of the five PEST residues.
Measurement of cell volumes Cells were grown in 10 ml of YEPD in a roller at 30°C to -.2 x 107 cells/ml. Culture tubes were then placed in ice, and then sonicated. Cells were diluted into Isoton buffer, and cell volume was analyzed using a Coulter Counter Model ZM (70 fim aperture), and a Coulter Channelyzer Model 256 calibrated with 5.11 lim diameter plastic beads (Coulter Electronics). Volumes reported in the text are mode volumes.
Propidium iodide staining, and measurement of length of G, phase Cells were grown as above, placed in ice and sonicated. I1 x 07 cells were harvested by centrifugation, and resuspended in 3 ml of water. 7 ml of 95 % ethanol was added slowly while the tube was vortexed. Cells were incubated in 70% ethanol overnight at 4°C. Cells were harvested, resuspended in 5 ml of 50 mM sodium citrate, pH 7, sonicated again and then harvested and resuspended in 1 ml of the same solution. RNase A was added to a final concentration of 0.25 mg/mil, and the cells were incubated at 50°C for I h. 1 ml of 50 mM sodium citrate 16 lkg/ml propidium iodide was added. After incubating for at least 30 min, the cells were filtered through a 36 itm mesh, and analyzed with a Coulter Model Epics-C Flow Cytometer. To quantify the GI peak, we first located the modes of the GI and the G2 peaks. The point midway between these modes was taken to be the midpoint of S phase. S phase was assumed to occupy 20% of the cell cycle, and so 10% of the total area under the curve was subtracted from the area to the left of the midpoint of S. Values reported in the text for GI were the remainder of the area left of the midpoint of S. The method's accuracy depends on whether or not S phase is truly 20% of the cycle; however, the method is objective and repeatable. a Factor resistance assays a Factor resistance was assayed in plastic grids, with each grid 1 cm square. Grid squares were individually filled with 1.5 ml of pH 4.5 YEPD + agar containing the appropriate concentration of a factor. After the agar had cooled and solidified, 5 11 of cell suspension containing 5 x 103 cells was spotted onto the surface. Results were scored after 2 - 3 days incubation. Quantitative matings were as described (Futcher and Carbon, 1986), but with a 1: 1 ratio of cells. Stationary phase analysis The proportion of cells in stationary phase was measured as described (Plesset et al., 1987), except that cells were incubated at 45°C instead of 48°C.
Acknowledgements We thank P.Sudbery, B.Gallay, R.Rothstein,W.Courchesne, J.Thorner, J.Boeke and D.Kaback for supplying strains and plasmids; C.Harley, M.Tyers, D.Visentin, C.Greider, J.Pringle, L.Molz, D.Beach, C. McLaughlin and A.Klar for helpful discussions; R.Booher and B.Byers for information concerning cdc13 and cyclins; C.Stojcic and M.Hazlett for technical assistance; L.Rodgers for flow cytometric analysis; G.Otto and C.Harley for help with the computer analysis; and B.Stillman, C.Greider and S.Cameron for comments on the manuscript. We especially thank F.Cross for stimulating discussions, and for communicating results prior to publication. This work was supported by the National Cancer Institute of Canada, and NIH grant GM39978.
References Bender,A. and Sprague,G.F., Jr (1987) Cell, 50, 681 -691. Bennetzen,J.L. and Hall,B.D. (1982) J. Biol. Chem., 257, 3026-3031. Booher,R. and Beach,D. (1987) EMBO J., 6, 3441-3447. Booher,R. and Beach,D. (1988) EMBO J., 7, 2321-2327. Calvert,G.R. and Dawes,I.A. (1984) Nature, 312. 61 -63. Carter,B.L.A. and Jagadish.N.N. (1978) Exp. Cell Res., 112, 373-383. Carter,B.L.A. and Sudbery,P.E. (1980) Genetics, 96, 561-566. Chan,R.K. and Otte,C.A. (1982a) Mol. Cell. Biol., 2, 11-20.
Chan,R.K. and Otte,C.A. (1982b) Mol. Cell. Biol., 2, 21-29. Coleman,K.G., Steensma,H.Y., Kaback,D.B. and Pringle,J.R. (1986) Mol. Cell. Biol., 6, 4516-4525. Dayhoff,M.O., Schwartz,R.M. and Orcutt,B.C. (1979) In Dayhoff,M.O. (ed.), Atlas of Protein Sequence and Structure. NBRF, Washington, DC, Vol. 5, suppl. 3, pp. 345-362. Dietzel,C. and Kurjan,J. (1987) Mol. Cell. Biol., 7, 4169-4177. Donnan,L. and John,P.C.L. (1983) Nature, 304, 630-633. Dumont,M.E., Ernst,J.F., Hampsey,D.M. and Sherman,F. (1987) EMBO J., 6, 235-241. Dunphy,W.G., Brizuela,L., Beach,D. and Newport,J. (1988) Cell, 54, 423 -431.
Evans,T., Rosenthal,E.T., Youngblom,J., Distel,J. and Hunt,T. (1983) Cell, 33, 389-396. Fantes,P.A. (1977) J. Cell Sci., 24, 51-67. Futcher,B. and Carbon,J. (1986) Mol. Cell. Biol., 6, 2213-2222. Gautier,J., Norbury,C., Lohka,M., Nurse,P. and Maller,J. (1988) Cell,
54, 433-439. Hamilton,R., Watanabe,C.K. and de Boer,H.A. (1987) Nucleic Acids Res., 15, 3581 -3593. Hartwell,L.H. and Unger,M.W. (1977) J. Cell Biol., 75, 422-435. Hartwell,L.H., Mortimer,R.K., Culotti,J. and Culotti,M. (1973) Genetics, 74, 267-286. Henikoff,S. (1984) Gene, 28, 351-359. Hereford,L.M., Osley,M.A., Ludwig,J.R., II and McLaughlin,C.S. (1981) Cell, 24, 367-375. Hill,J.E., Myers,A.M., Koerner,T.J. and Tzagoloff,A. (1986) Yeast, 2, 163-167.
Hindley,J. and Phear,G.A. (1984) Gene, 31, 129-134. Holm,C., Meeks-Wagner,D.W., Fangman,W.L. and Botstein,D. (1986) Gene, 42, 169-173. Ito,H., Fukuda,Y., Murata,K. and Kimura,A. (1983) J. Bacteriol., 153, 163- 168.
Jagadish,M,N. and Carter,B.L.A. (1977) Nature, 269, 145-147. Jenness,D.D., Burkholer,A.C. and Hartwell,L.H. (1986) Mol. Cell. Biol., 6, 318-320. Johnston,G.C., Pringle,J.R. and Hartwell,L.H. (1977) Exp. Cell Res., 105, 79-98.
Kelly,D.E., Trevethick,J., Mountain,H. and Sudbery,P.E. (1988) Gene, 66, 205-213. Killander,D. and Zetterberg,A. (1965) Exp. Cell Res., 40, 12-20. Kozak,M. (1986) Cell, 44, 283-292. Liskay,R.M. (1977) Proc. Natl. Acad. Sci. USA, 74, 1622-1625. Lord,P.G. and Wheals,A.E. (1980) J. Bacteriol., 142, 808-818. MacQueen,H.A. and Johnson,M.H. (1983) J. Embryol. Exp. Morphol., 77, 297-308. Maniatis,T., Fritsch,E.F. and Sambrook,J. (1982) Molecular Cloning. A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Mendenhall,M.D., Jones,C.A. and Reed,S.I. (1987) Cell, 50, 927-935. Miyamoto,S., Ohya,Y., Ohsumi,Y. and Anraku,Y. (1987) Gene, 54, 125-132.
Morgan,A.R., Lee,J.S., Pulleyblank,D.E. and Evans,D.H. (1979) Nucleic Acids Res., 7, 547. Nurse,P. (1975) Nature, 256, 547-551. Nurse,P. (1985) Trends Genet., 1, 51-55. Nurse,P. and Bissett,Y. (1981) Nature, 292, 558-560. Nurse,P. and Thuriaux,P. (1977) Exp. Cell. Res., 107, 365-375. Nurse,P. and Thuriaux,P. (1980) Genetics, 96, 627-637. Otto,G. (1986) Ph.D. Thesis, Massachusetts Institute of Technology. Pardee,A.B., Dubrow,R., Hamlin,J.L. and Kletzien,R.F. (1978) Annu. Rev. Biochem., 47, 715 -750. Piggott,J.A., Rai,R. and Carter,B.L.A. (1982) Nature, 298, 391-394. Pines,J. and Hunt,T. (1987) EMBO J., 6, 2987-2995. Plesset,J., Ludwig,J.R., Cox,B.S. and McLaughlin,C.S. (1987) J. Bacteriol., 169, 779-784. Prescott,D.M. (1956) Exp. Cell Res., 11, 86-98. Pringle,J.R. and Hartwell,L.H. (1981) In Strathern,J., Jones,E. and Broach,J. (eds) The Molecular Biology of the Yeast Saccharomvces Vol. 1. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Reed,S.I. (1980) Genetics, 95, 561 -577. Reed,S.I., Hadwiger,J.A. and Lorincz,A.T. (1985) Proc. Natl. Acad. Sci. USA, 82, 4055-4059. Robbins,E. and Scharff,M.D. (1967) J. Cell Biol., 34, 684-686. Rogers,S., Wells,R. and Rechsteiner,M. (1986) Science, 234, 364-368. Rothstein,R.J. (1983) Methods Enzymol., 101, 202-211. Russell,P. and Nurse,P. (1987) Cell, 49, 559-567. Sanger,F., Nicklen,S. and Coulson,A.R. (1977) Proc. Natl. Acad. Sci. USA,
4345
R.Nash et al.
74, 5463-5467. Shields,R., Brooks,R.F., Riddle,P.N., Capellaro,D.F. and Delia,D. (1978) Cell, 15, 469-474. Singer,R.A. and Johnston,G.C. (1981) Proc. Natl. Acad. Sci. USA, 78, 3030-3033. Sudbery,P.E., Goodey,A.R. and Carter,B.L.A. (1980) Nature, 288, 401-404. Swenson,K.I., Farrell,K.M. and Ruderman,J.V. (1986) Cell, 47, 861 -870. Trueheart,J., Boeke,J.D. and Fink,G.R. (1987) Mol. Cell. Biol., 7, 2316-2328. Van Arsdell,S.W., Stetler,G.L. and Thorner,J. (1987) Mol. Cell. Biol., 7, 749-759. Vieira,J. and Messing,J. (1987) Methods Enzymol., 153, 3-1 1. Yen,A., Fried,J., Kitahara,T., Strife,A. and Clarkson,B.D. (1975) Exp. Cell Res., 95, 295-302.
Received on September 23, 1988
4346
